Home - About Us - Riding Safer - Contact Us - Blog - Disclaimer - Links - Sitemap
Risk Hierarchy: Information - Rider Ed - Driver Ed - Conspicuity - Bike Defect - Ultra-Defensive Riding - Crash Avoidance - Injury Mitigation - Crash Scene

helmetHelmets: Do they work?

The research on the subject, as we point out in our statement on the helmet law issue, has been contaminated by research commissioned by NHTSA and other advocates of helmet laws, which are alleged to have issues with possible bias or presumptions about helmet use.

The current NHTSA numbers, by the way, use 2003 to 2005 data from 18 states.

Someone dies or goes to hospital in 40% of motorcycle crashes.

We think this is significant and probably accurate.

An Uncontaminated Helmet Study

We think that this paper, by Jim Ouellet and Vira Kasantikul, which looks at helmet use and outcomes in the original Hurt study and in the Thailand, might be relatively free of bias. We note Ouellet's affiliation with the Head Protection Lab., which performs helmet testing and is a proponent of helmet use. He worked on the original Hurt study. Dr. Kasanthul is a neurologist and has done previous work on motorcycle crash head trauma. These are very highly respected experts in their field and were following the OECD guidelines, which are free of obvious helmet bias.

We have some of their findings in the following table.

Helmet Use with Casualty Figures (Helmets retained)
Los Angeles 1981
Thailand Study 1999-2000
No Helmet
No Helmet
9 (3.4%)
21 (7.6%)
8 (1.9%)
21 (6.8%)
Brain Injury
27 (7.9%
67 (13.5%)
8 (1.9%)
57 (11.6%)

Severe brain injury

5 (1.5%)
24 (4.8%)
2 (0.5%)
10 (2.0%)

Notes about Statistics:

The following section provides infromation about the statistical reliability of the numbers in the table.

In Los Angeles, unhelmeted riders were more than twice as likely to be killed as helmeted riders on a per-accident basis (7.6% vs. 3.4%) and the difference was statistically significant (χ2 = 6.92, p = .008, df = 1). Helmeted riders were 40% of the accident sample but only 23% of the fatalities.

In Thailand, unhelmeted riders were more than three times as likely to be killed, and helmet-ejected riders were more than six times as likely to be killed as riders whose helmet stayed on. That is, only 1.9% of the helmet-retained riders, compared to 6.8% of the unhelmeted riders died and 12.8% of helmet-ejected riders. The difference among the three groups was statistically significant (χ2 = 25.05, p < .001, df = 2) and all three pair-wise comparisons were significantly different (χ2 > 5, p < .02, df = 1).

In Los Angeles, a total of 94 of the 839 riders who survived (11%) suffered some level of brain injury (AIS 1-5), and 29 riders (3.5%) had a brain injury in the AIS 3-5 (serious-to-critical) range. However, brain injury rates differed as a function of helmet use. Unhelmeted surviving riders suffered a brain injury at any level of severity (AIS 1-5) nearly 60% more often than helmeted riders (13.5% vs. 7.9%) – a difference that was statistically significant (χ2 = 6.5, p = .011, df =1). Unhelmeted riders were also three times as likely (4.8% vs. 1.5%) to suffer a major brain injury (AIS 3-5). Comparisons of brain injury rates among surviving riders are shown...

Among the 1019 Thai riders who survived, unhelmeted riders were nearly six times as likely as helmet-retained riders (11.6% vs. 1.9%) to suffer a brain injury at any level of severity (AIS 1-5). Helmet-ejected riders who survived fared even worse, with 26% sustaining some kind of brain injury. The differences were statistically significant, and all three pair-wise comparisons differed significantly (χ2 > 15.6, p < .0001, df =1). Only two of the 411 helmet-retained surviving riders (0.5%) suffered AIS 3-5 brain injury, compared to 2.0% of 492 unhelmeted survivors and 2.6% of survivors whose helmet ejected. Although the overall difference in the frequency of AIS 3-5 brain injuries was not significantly different (p = .09), helmet-retained survivors had significantly fewer AIS 3-5 brain injuries than unhelmeted riders (χ2 = 4.08, p = .043, df = 1). Comparisons involving helmet-ejected riders were not statistically significant.

(The above quoted from "Motorcycle Helmet Effect on a Per-Crash Basis in the Thailand and Hurt Studies", Ouellet and Kasantikul. with thanks.

Our comment on Hurt/Thailand numbers

The table says it all: Helmet users are much less likely to die, three or four times less likely to be vegetative or severely disabled due to head trauma, and significantly less likely to have treatment for brain injury. Helmet use reduces head injuries significantly and saves lives. These numbers were mostly statistically significant, with the exception of the serious brain injury, where the numbers were small. The methodology used is the best available, there is no possibility of fakery, and there is no shadow of a doubt that they are good.

We note that only helmets that stayed on are counted for helmet use. A helmet that comes off in a crash, seems to make you much more likely to suffer injury or death. It seems worse than no helmet at all.

Thai helmet standards are more lax than the US DOT standard. US helmet internal shells are more than 20% thicker on average, and we'd expect them to work better.

One word of caution might be that Thai motorcycles are more likely to be scooters, mopeds or smaller-displacement lighter bikes than the typical US motorcycle, which is capable of higher speeds and higher-energy crashes. We can't expect this level of protection in US crashes, and even the helmet numbers claimed by NHTSA don't claim Thai levels of efficacy.

Helmets that Came Off

The Thai study shows an increased risk of injury where a helmet was worn but came off during the crash.

There were 111 cases of helmets coming off. Of these, twelve were due to some type of damage during the crash, and the remaining 98 were some combination of helmet not fastened, helmet size too large or fastened too looely.

In the remaining 13 cases, two were caused by the riders jaw being broken and unable to keep the strap in position. The remaining 11 cases were some failure of the helmet shell or retention system that arose during the crash.

If you take the view that a full-face helmet would probably have prevented the jaw injuries, there remain nine cases where the helmet failed and came off during the crash, and the rider had some level of injury and death in addition to those listed in the table above. We don't have a breakdown of those numbers, but we feel that some of them would have been shown in the 'helmet' death and injury columns, and, given the nature of the helmet retention failure happening in a very violent crash, we'd expect all 9 to show up in the killed or brain injury columns.

If this was so, we'd expect to see the death rate for helmeted riders at about 40% and the serious brain injury at about one third the level of helmetless riders.

Hurt didn't say much about helmets that ejected, though we'd expect a similar outcome.

Neck Injuries

The Thai study also reported neck injuries in detail. In the Bangkok part of the study, there were 42 riders with cervical spine fractures or dislocations compared with a total of 261 head injuries. Although 41 of these riders died, it seems that other injuries were the cause of death in most cases. The neck injuries were mostly in combination with other massive injuries. The report lists spinal injuries in total as the cause of death in three rider cases. Dr Kasantikul is a neurologist and performed a special autopsy procedure on the head and neck of all but three of the dead riders. He found a total of 109 soft tissue neck injuries, which would not usually be found in autopsies. Man of these injuries were not found by the trauma physicians. This leads, in his view, to a severe underreporting of neck injuries in general in bike crashes.

The maximum numbers, 2 or 3 fatal neck injuries and 42 neck spinal injuries are small in comparison to the 261 head injuries, and the 60 rider deaths from all causes.

We see no reason to deduce that the wearing of helmets led to a large increase in neck injuries. In view of the 265 riders whose head injuries were reduced or prevented by the use of the helmet, the scope for any collateral damage from helmet wearing must be small.

We should anticipate that neck injuries might be under-reported in other studies and statistics because pathologists are not trained to perform special neck inspections for bike fatalities.

Maids findings on helmet use:

Maids was done in Europe, where helmet usage is mandatory in all the areas studied. There is no helmet law controversy in Europe to any significant extent. 8% of riders in the accident sample were not wearing helmets, all except 4 were riding mopeds or scooters. 67% of the helmeted riders were wearing full-face helmets, more than we'd expect in the US.

As bikers skulls, concrete and the laws of gravity are the same in Europe as the US, a look at these findings is interesting.

These numbers were developed by comparing helmet impact damage with the head injuries. The helmets that had no effect on brain injury were mainly torsion rather than impact injuries.

Interestingly, 73 (8.0%) of the riders in crashes were not wearing helmets, and 70 (7.6%) of the population sample, indicating that not wearing a helmet does not have any significant effect on the ability to avoid crashes.

Only 5% of the crash riders had spinal injuries, including neck spinal injuries. There's no real evidence that 90% of riders who were wearing helmets suffered additional neck injuries. The small number of overall spinal injuries, are far outweighed by the 68.7% of riders who had reduced head injuries because of their helmet use.

In terms of bias, the Maids study used the OECD methodology, which was largely developed by Head Protection Research Lab. people, and had Terry Smith of Dynamic Research and Head Protection Research Lab on staff as a consultant.


Here's a report put together by a non-helmet-wearing biker who works for the MN Dept of Public Safety. It shows 11 (18% of total deaths) helmeted riders in 2007, and 12 (16.7%) in 2007 died in crashes. 45 (73.8%) and 53 (73.6%) of riders who were killed wore no helmets in those years. Total deaths were 61 in 2007 and 72 in 2008.

59.1% of Minnesota riders self-report wearing a helmet 'all or most of the time'. There was no control group for the study, and these self-reported numbers tend to overstate, so the commonly accepted rule of thumb for most states several years after rescinding helmet laws of about 50% is probably more accurate.

Assuming 50% helmet use, it looks like not wearing a helmet in Minnesota gets you killed at about four times the rate of helmeted riders. The overall numbers of injuries for helmeted and unhelmeted riders were about the same, roughly 70% of both groups. There were 1633 crashes reported in 2008.

There are no statistical measures of significance provided with these numbers, but our guess, considering the number of deaths (61 and 72), is that they would probably be significant.

The author of the study self-identified as a non-helmet wearing biker to us when he sent the link.


As we'll not likely get much information in helmet efficacy from the upcoming OSU crash causation study, we'll have to settle for information like this.

In both of the papers discussed above, we pointed out previous statements from some of the authors on the efficacy of helmets.

We can't reject these findings out of hand. These are serious, ethical and learned researchers, and have done a lot of research over the years on which to base their opinions. They have done enough work to earn a fair hearing.